Security News

Black Friday Brings AI Shopping Agents and a Fresh Wave of Scams, Arctic Wolf Warns

Key Takeaway Keep security on the priority list during peak trading by adding extra checks for high-risk purchases and watching for unusual activity on staff devices and customer accounts.

Arctic Wolf has warned Australian retailers and consumers to prepare for a rise in fraud, impersonation and AI-enabled scams during the Black Friday shopping period. The company says the pressure is greatest on smaller businesses, where every sale counts and security can quickly fall down the priority list. It pointed to the recent voluntary administration of David Jones as a sign of a difficult trading environment.

The warning is backed by figures. Fortinet FortiGuard Labs found more than 18,000 holiday-themed domains registered in the three months before the 2025 holiday season, with at least 750 confirmed as malicious. Australia's National Anti-Scam Centre reported $158.5 million in losses from scams started via websites, mobile apps and social media in 2025, up 21.8% on the previous year.

Arctic Wolf expects five threats to feature heavily: fake retailers and brand impersonation tied to payment fraud; "verification" prompts designed to deliver malware; account takeovers and information theft; AI-assisted scams that are harder to spot; and personal AI shopping agents that could be tricked into buying from fraudulent sellers. The malware technique, known as ClickFix, uses fake CAPTCHA tests or browser errors to trick users into running malicious commands. Microsoft has reported it targeting thousands of devices globally each day. Stolen browser cookies and authentication tokens can also let attackers hijack sessions, in some cases bypassing password-based controls.

For retailers, Arctic Wolf recommends controlling which AI shopping agents can transact, separating verified agents from unidentified automation and malicious bots, and requiring step-up authentication for gift cards, loyalty redemptions and large purchases. It also advises monitoring for browsers launching PowerShell or command line tools, and giving AI agents only the permissions they need while treating reviews, emails and uploaded documents as untrusted.

Why it matters

Retailers can no longer assume that a fast, well-formed purchase comes from a human customer, because authorised AI agents and malicious bots will look alike unless they are identified. Stolen session tokens also break the assumption that a password and login check protect an account, which makes step-up checks on gift cards, loyalty points and large orders a real decision rather than a nice-to-have. Small retailers with thin security staffing during the sales rush are the most exposed.

What to do

  • Require step-up authentication for gift card purchases, loyalty redemptions and large orders.
  • Monitor for browsers launching PowerShell or command line tools, which can indicate ClickFix activity.
  • Separate verified AI shopping agents from unidentified automation and malicious bots.
  • Limit AI agent permissions to what is necessary and treat reviews, emails and uploaded documents as untrusted.
Running AI agents? Eight controls for securing them, mapped to the ISM ->

Summarised by CISO AI from Australian Cyber Security Magazine, written with Claude Sonnet 5.5. We link back to every original so you can read it yourself.